AI Prompt Engineering and Key Concepts in Machine Learning and NLP Practice Test

These questions are part of the practice quiz. Start practicing

• How does tokenization impact model input constraints and prompt design in practice?
• Explain the concept of FEVER-style evaluation in the context of factuality assessment.
• What term describes the process by which AI systems acquire knowledge, refine their algorithms, and improve their performance over time based on input data, including detailed prompts?
• Which term best describes the process of generating a concise version of a document while preserving its main ideas?
• Which term denotes the quality of maintaining uniformity and coherence in behavior across interactions?
• Which term is defined as 'a parameter that influences randomness in generated outputs'?
• Which method uses a small number of examples to guide the model's responses?
• Which term describes the extent to which a product or system can be used effectively, efficiently, and satisfactorily by users to achieve their goals?
• Which term covers the moral principles and guidelines governing AI development to promote fairness, transparency, accountability, and societal well-being?
• Which term describes clear and specific instructions provided to AI systems to guide their responses?
• To reduce reliance on prompt surface cues, which practice is advisable?
• What term describes the equitable treatment of individuals across demographic groups in AI systems to prevent unfair outcomes?
• What is data augmentation in prompts and what effect does it have?
• What is the primary benefit of self-consistency in chain-of-thought prompts?
• Which concept is about ensuring AI systems are not biased toward a particular group due to faulty data?
• Which concept is most relevant for evaluating the stability and dependability of AI outputs across interactions?
• How do backpropagation and gradient descent relate to prompt tuning vs full-model fine-tuning?
• What are adapters and LoRA in the context of efficient fine-tuning?
• Which statement best describes the purpose and best practices of background knowledge retrieval in prompts?
• What is semantic similarity and how is it used in prompt retrieval?
• Which term describes the level of detail, realism, and artistic style of the images produced by AI image generation tools?
• Which term corresponds to the definition: 'a user query, command, or input in an AI interface'?
• What does the RAG architecture consist of, and how do the components function together?
• What is data leakage in ML experiments and how can it affect evaluation?
• Explain ROUGE metrics and their typical use in summarization; what are their limitations?
• Explain attention mechanism and why it enables transformer models to capture long-range dependencies.
• Which term matches 'refers to instances where the model correctly identifies inputs or conditions as not meeting certain criteria or expectations'?
• Which NLP task involves identifying and classifying named entities such as persons, organizations, locations, and dates within text?
• Which term describes features allowing users to adjust and refine generated images, including iterative feedback and style variations?
• What are safety alignments and rule-based prompting used for?
• What is a recommended approach to preserve essential context in a constrained memory setting for a chat prompt?
• Which term describes 'extra data provided to add context or background information'?
• Which term matches 'refers to a fictional character or user profile created to represent a specific demographic, behavior pattern, or set of characteristics'?
• What is prompt chaining?
• Which term describes 'the information provided to a system as input for processing'?
• Which term refers to the correctness and precision of the information contained within a dataset, crucial for reliable AI predictions?
• Which concept would be most relevant when an organization wants to help customers understand how their data influences automated decisions?
• Which term refers to a lack of clarity or specificity in prompts, leading to potential confusion?
• Which term refers to turning printed or handwritten text into machine-readable text?
• Define prompt diversity and explain why it matters for robust evaluation.
• Which term matches 'supplementary data or details provided alongside the main content to provide context, clarification, or background information'?
• How can you measure factuality in LLM outputs, and what evaluation approaches are commonly used?
• Which aspect best supports robust evaluation across prompts?
• In chain-of-thought prompting, self-consistency refers to
• Which NLP task labels each word with its grammatical category such as noun or verb?
• What is the process of identifying and correcting errors, inconsistencies, and inaccuracies in datasets to ensure data quality?
• Which prompt engineering technique offers a limited number of examples or shots related to a specific task or topic to guide the language model's response generation?
• What term refers to the raw information used to train, validate, and test machine learning models?
• Which term encompasses data's completeness, consistency, and reliability through its life cycle and guards against tampering to preserve trustworthiness?
• What term describes condensing a software program or codebase into a shorter form while preserving functionality and logic?
• Which element is essential in a robust evaluation plan for prompt-based tasks?
• Which term refers to the surrounding circumstances that influence the interpretation or meaning of information?
• Which prompt engineering technique provides input to a language model without specific training examples, expecting it to generate a response based on its preexisting knowledge?
• Which NLP task is used to automatically pull out dates, names, and locations from text?
• Which term best matches the goal of ensuring AI decisions do not systematically disadvantage any group?
• Which phrase best describes the different fields or industries where AI systems are applied, such as healthcare, customer service, education, and finance, each with unique requirements and challenges in AI interaction?
• Which prompt engineering technique starts with simple prompts and gradually increases complexity based on the AI's responses to guide the AI effectively while minimizing user effort?
• What is the process of using AI to forecast future outcomes based on historical data called?
• Which term describes condensing a codebase into a shorter representation while retaining its core behavior?
• In Retrieval-Augmented Generation (RAG), what are the two main components and their roles?
• Which term matches 'the information or data provided to a system, model, or application as input for processing or analysis'?
• What is a key limitation of ROUGE metrics?
• Describe zero-shot chain-of-thought and how it compares to standard chain-of-thought prompting.
• Which statement about the Brier score is correct?
• Which statement about temperature and top-p control is accurate?
• Which prompting technique encourages LLMs to explore different possibilities and consider multiple reasoning paths before generating a response?
• Which technique involves prompting the model to show its reasoning steps?
• Which statement correctly describes how to mitigate prompt gaps and leakage?
• Which term refers to the moral standards governing AI to ensure fairness, transparency, accountability, and respect for human rights?
• How can you mitigate prompt gaps and leakage?
• Which term contrasts with misinformation by being deliberately deceptive?
• What is calibration in the context of probabilistic outputs from language models, and how can you calibrate them?
• What is LoRA and how does it help in fine-tuning models?
• Which technique is associated with generating a sequence of reasoning steps, often called chain-of-thought?
• Explain what makes an evaluation dataset effective for NLP prompts.
• Explain encoder-only, decoder-only, and encoder-decoder transformer architectures and typical NLP tasks for which each is used.
• Which term best captures the idea of refining and steering generated images through repeated feedback and parameter adjustments?
• In a confusion matrix for a multi-class NLP classifier, what do the diagonal entries represent?
• Which term describes 'the data structure and visual arrangement of results'?
• What is calibration error and how would you evaluate alignment of confidence with accuracy in a prompt-based task?
• What is the initial step in designing and developing new software applications, artistic content, and prototypes using generative AI?
• Which term refers to the field that enables computers to learn from data and make predictions or decisions?
• What is the initial step in data analysis where raw data is cleaned, formatted, and transformed to prepare it for study?
• When discussing AI deployment across healthcare, finance, and education, which term describes these contexts collectively?
• Which term describes the presence of unfair or prejudiced outcomes in AI systems due to biases in data, algorithms, or design?
• What term refers to the emotional or expressive quality conveyed in written text?
• Which concept focuses on creating visual representations to reveal patterns and insights in data?
• Which statement correctly contrasts instruction-tuning and prompt-tuning?
• Which category includes techniques such as few-shot, zero-shot, tree-of-thought (ToT), chain-of-thought (CoT), and self-consistency?
• Which term matches the definition 'occurs when the model incorrectly generates outputs that are not relevant or accurate to the input or task'?
• What does a confusion matrix represent in NLP classification tasks, and which metrics are commonly derived for imbalanced data?
• Which term matches 'the structure, layout, and presentation style of the results or outputs generated by a system, model, or process'?
• If you want an AI system to understand and align with the user's underlying goal behind a prompt, which concept is most relevant?
• Which term is used to describe forecasting future outcomes based on historical data?
• What is the main advantage of Retrieval-Augmented Generation (RAG)?
• Which concept refers to the underlying purpose or goal behind user queries or prompts, which AI systems strive to understand and address accurately?
• Which term corresponds to the definition: 'the means by which users interact with and provide input to AI models to generate desired outputs, such as text, images, or other forms of content'?
• Which prompt engineering technique is used to encourage large language models to explain their reasoning by revealing the intermediate steps the model takes to arrive at an answer?
• In the context of prompts, how can data leakage occur?
• Which term describes the activity of removing duplicates and handling missing values to prepare data for modeling?
• What feature involves user interaction to refine and improve generated images?
• What is the process of categorizing data into predefined classes based on input features using a machine learning algorithm?
• Which statement best describes the bias-variance tradeoff and its impact on model generalization in NLP?
• What is the difference between generative search and traditional search in information retrieval?
• Which of the following is a direct effect of system messages on prompt design?
• Which architecture is best suited for translation and summarization tasks?
• What is the primary role of system messages in prompt design?
• What term describes the set of rules that constrain the behavior or design of a system?
• Which practice would most directly support accountability by making AI decision processes auditable?
• Which term refers to descriptive keywords or parameters like mood, lighting, viewpoint, or style?
• Which is a common safety-oriented constraint used in rule-based prompting?
• Which NLP task involves assigning grammatical categories or tags to each word in a sentence based on its syntactic function?
• What is the process of condensing a longer text into a shorter version while keeping key ideas?
• Which statement about BLEU is correct?
• Which term refers to the quality of AI responses that accurately address the specific requirements or queries, minimizing errors?
• Which term refers to clear and specific instructions provided to AI to guide responses, often including context and constraints?
• Which practice is recommended for handling model drift in prompt maintenance?
• What are confounding variables in human evaluation of prompts and how can you control for them?
• Which concept describes AI systems' ability to understand the background, nuances, and requirements of user queries to generate contextually appropriate responses?
• Which prompting technique involves a two-step process where initial knowledge is generated and then used to inform the main prompt?
• What is dataset shift and how can it affect prompt generalization across domains?
• Which term captures the capacity of AI to adjust its behavior in response to changing contexts or user preferences?
• Which subset of AI enables computers to learn from data and make predictions or decisions without explicit programming?
• Which term describes the need to explain the rationale behind AI decisions to stakeholders?
• What is a token in NLP and how does tokenization affect prompt length and model input constraints?
• Which term focuses on data's completeness, consistency, and reliability throughout its lifecycle, including protection against tampering?
• Which term captures users' confidence in the accuracy, consistency, and usefulness of AI-generated responses built through reliable and relevant interactions over time?
• Which process involves refining prompts through successive iterations by adding additional details or modifiers?
• Which term corresponds to the degree to which information or responses directly address the user's query, needs, or context?
• Which prompt engineering technique involves the model asking itself additional questions to gain a deeper understanding of the user's prompt before formulating a response?
• Which term describes false or inaccurate information generated by AI unintentionally due to training data and algorithm limitations, contrasting with deliberate deception?
• Which statement best describes zero-shot prompting in language models?
• Which attribute of AI ethics emphasizes stakeholder understanding of how decisions are made?
• Define perplexity and explain its role as a language model evaluation metric.
• What is embedding in NLP, and how do contextual embeddings differ from static embeddings?
• Which term refers to using AI to automate video editing tasks like scene detection, caption generation, and music selection?
• Which is a common failure mode of prompt-based systems?
• Which factor is most directly influenced by system messages during prompting?
• What is the graphical representation of data and information used to communicate insights?
• How would you design a prompt to handle multi-turn dialogue with memory constraints?
• What is cross-entropy loss and how does it differ from hinge loss?
• What is a BLEU score and what are its limitations for evaluating natural language generation?
• Which term refers to safeguards designed to prevent discriminative outcomes in AI predictions?
• Compare BERT-style encoders with GPT-style decoders in terms of architecture and typical use cases.
• Which issue arises when AI models perform differently across groups because of data biases?
• How can prompt injection pose security risks, and what mitigations exist?
• What is the process of automatically identifying and extracting specific pieces of text or information from a larger document or source?
• Which term describes the various ways users can input to AI image generation tools, such as textual descriptions, sketches, and prompts?
• What is soft prompts and how do they enable prompt tuning without updating the base model?
• Which term is most associated with auditing AI decisions for errors and discrimination?
• Which term describes the AI-driven technology that converts printed or handwritten text into machine-readable format?
• Which term describes the absence of precision or specificity in prompts, allowing multiple interpretations?
• Which statement about chain-of-thought prompting is correct?
• Which term best describes the broader practice of building predictive models from data using algorithms?
• What is retrieval-augmented prompt design best practice and how to ensure citations for facts?
• Which statement best describes prompt engineering compared to fine-tuning?
• Which phrase refers to systematic errors in AI that produce unfair results due to biased data or design?
• Which term refers to descriptive attributes added to prompts to influence mood, lighting, or viewpoint?
• Which NLP task is used to categorize text documents into predefined topics or categories based on content?
• Which prompting technique asks the model the same prompt multiple times and takes the most consistent result as the final answer?
• In prompting, how does prompt complexity influence generalization to unseen prompts?
• The phenomenon of discriminatory AI results caused by biased training data is known as what?
• In imbalanced NLP classification tasks, why use macro or micro averages in confusion-matrix-based metrics?
• Which AI-driven technology transcribes spoken words into text?
• Which NLP task involves categorizing text documents or instances into predefined classes or categories based on their content?
• Which term is defined as 'a parameter in the LLM that influences the level of randomness or creativity in generated outputs'?
• Which term denotes the ease with which users can operate a product to achieve goals?
• Which concept aims to enhance trust and reliability by explaining AI decisions?
• Which term denotes the practice of making AI decision processes understandable to users?
• How does prompt length interact with the attention budget, and what design strategies help?
• What is model drift and its implications for prompt maintenance?
• Which term describes the specialized language used within specific fields or communities, often unfamiliar to those outside of that context?
• Which metrics are commonly used to evaluate generation tasks in prompt-based systems?
• What is the purpose of AUROC, and when is it particularly useful?
• Which term describes the ability of AI to adjust its responses based on changing contexts or user preferences?
• Which term is defined as the quality of being easily understood and free from ambiguity in communication?
• What is instruction following evaluation and how is it measured in prompt-tuned models?
• What is dataset leakage in evaluation sets, and why is it problematic?
• How do transfer learning and multi-task learning differ, with NLP examples?
• What describes statements or answers that provide accurate information based on verifiable facts?
• Explain token budget management in chat prompts and strategies to maximize information within it.
• Which concept addresses lack of clarity in how AI outputs are produced, potentially undermining user trust?
• What process identifies and confirms the speaker's identity for secure authentication?
• What term denotes the level of detail or brevity in the language of a prompt affecting information communicated to the AI?
• Which term describes detailed instructions that specify image type, main subject, background, and composition style?
• Which concept focuses on making AI decisions understandable to users and stakeholders, promoting trust and accountability?
• Which term refers to 'a fictional character used to represent a demographic'?
• What term describes the ethical and accountable development, deployment, and use of AI systems that prioritizes fairness, transparency, accountability, safety, and societal well‑being while mitigating risks?
• Which term denotes the overall quality of interactions between users and AI systems, including ease of use, satisfaction, and effectiveness?
• In prompt design, what practice helps prevent hallucinations when using background knowledge retrieval?
• Which term is defined as 'occurs when the model outputs are not relevant or accurate'?
• Which set of prompting methods is typically described as including few-shot and zero-shot prompting?
• Which term best describes the idea of maintaining uniform tone and behavior across many interactions?